The graveyard economy: why a billion dollars of side project code is sitting unused

Somewhere on a private GitHub repo right now, there is a SaaS app that took eight months to build, has not been touched in four years, and would save the next builder roughly a year of work. The founder will never look at it again. It will never run for a paying customer. It will quietly age out as the framework versions drift, the API keys expire, and the .env file becomes archaeology.

Multiply that by every indie developer you know. Then by every indie developer you do not know.

That is the graveyard economy. I think it is worth more than a billion dollars in the aggregate, and I think the reason almost none of that value gets unlocked is one of the more interesting market failures of the post ChatGPT software era.

The math, with reasonable numbers

Let me show my work, because the headline number sounds like linkbait until you do the arithmetic.

There are roughly five million people in the world who would self identify as an “indie developer” in 2026. Some estimates from GitHub and Stack Overflow data put the figure higher. Five million is a conservative floor.

Most of these developers have shipped or partly shipped more than one side project. In conversations and surveys, the average sits around three. Some have ten in a folder. Some have one. Three is the rough middle.

Now the replacement cost. A working SaaS MVP in 2026 costs between $20k and $50k to rebuild, whether you measure it in freelance dev hours, agency quotes, or the opportunity cost of the founder’s own labor. I covered the valuation framework in another piece. That number lines up with what buyers and sellers on Failedups actually agree on once they get past the sunk cost arguments.

So the unadjusted ceiling is five million developers, times three projects each, times $30k average replacement cost. That is $450 billion. Obviously absurd, because most of those projects are not commercially viable, most are duplicative, and most have rotted.

Apply the heaviest discount you can defend. Say only one in a hundred is actually transferable, runnable, and has a coherent enough thesis that someone else could pick it up. A 99 percent haircut still lands at $4.5 billion of dormant value on private repos.

Cut it again. Say the realistic clearing price is a tenth of replacement cost, because buyers are not paying full freight for a stranger’s half built thing. You are still at $450 million. Tighten the assumptions in any honest direction and the number lives in the high hundreds of millions to low single digit billions.

Whatever your discount, the point is the same. There is a real, large, and almost completely illiquid pool of code out there. The people who built it are getting nothing for it.

Four reasons the market does not clear

If a billion dollars of value were sitting in any other asset class, capital would have already shown up. Why has nobody built the stock exchange of unfinished SaaS? I think there are four reasons, and they stack.

The first is shame. Founders use the word “kill” when they talk about ending a project. They do not say “list” or “transfer.” Killing implies a clean death. Listing implies an admission, in public, that you are not the right person to finish what you started. That is socially expensive in a culture where the founder narrative is “and then I just kept going.” Most people would rather let the project rot than write the for sale post.

The second is liquidity. Until very recently, there was no equivalent of a property market for stalled SaaS. You could try Acquire, but Acquire is built for revenue. You could DM around your network, but your network does not contain enough buyers for a $500 to $10k asset. Even the founders who do want to sell run into the same wall: who would I even tell.

The third is the technical complexity of transfer. Stripe accounts have to migrate. Domains have to transfer. The .env file has to be cleaned of secrets without breaking the app. Documentation has to be written from memory, which is the worst kind. None of this is hard, but all of it is annoying, and annoying friction at the bottom of a price curve kills deals at scale.

The fourth is a misaligned mental model of value. Founders tend to price their stalled project in one of two ways. Either it is worthless, because they never made it work. Or it is priceless, because they spent their own time on it, and their time is worth their salary, so the project must be worth six figures. The honest middle, where a real buyer might pay $3k for something that took $25k of labor to produce, is psychologically uncomfortable for many sellers. So they hold, they wait, and the code rots.

What this looks like at the human scale

A founder I spoke with last month built an analytics tool for Shopify merchants. Eight months of nights and weekends, a working dashboard, three paying beta users at $29 a month, a domain she still owns. New job came up. The project went into the freezer. Four years later, she still pays $14 a year for the domain, more out of guilt than intent.

Someone with marketing chops and twelve free hours a week could acquire that project, refresh the framework, and be twelve months ahead of where they would be building from scratch. She gets $4k and closure. They get a launching pad. That trade does not happen, today, mostly because there is no path between her DM and that buyer. The market does not clear. The value sits.

Imagine that situation playing out a hundred thousand times across a hundred thousand private repos. That is the economy I am describing.

The buying-and-improving decade

Here is the part I think is genuinely interesting, and the part I have come to believe more strongly the more listings we read.

The next decade of indie business is going to look less like a generation of greenfield builders and more like a generation of buyers who flip and improve. The same way the housing market evolved.

Greenfield is the dominant mode today because the cost of starting a project is almost zero. AI tooling makes the first weekend exhilarating. The problem is that the first weekend is the easy part. The slog from “demo works” to “first ten paying customers” is where projects stack up in the graveyard. That slog has not gotten meaningfully cheaper. Distribution is still expensive, and patience is the most expensive thing of all.

If you accept that, the case for buying gets stronger every year. The same way it became more rational at some point to buy and renovate a house in a good neighborhood than to build one from scratch on raw land. The land is mostly taken. The interesting move is the renovation.

A buyer who acquires a half built SaaS for $3k is not getting a free company. They are getting a foundation. The walls are up. The plumbing mostly works. They still have to do the kitchen and convince anyone the place is worth living in. That is the hard part. But they are not laying the slab.

I think we are at the early edge of this turning. The graveyard is getting bigger every year, and a generation of builders is learning that founding does not have to mean originating. Sometimes the best move is to take the keys.

That is what we are trying to build at Failedups. A liquid, honest, low friction market for the half built thing in your folder. A way to convert eight months of dormant work into either four thousand dollars or a partner who actually finishes it. A path that is neither “pretend it never happened” nor “guard it forever in case I come back.”

If you have a project in the freezer, you can list it on Failedups and find out what it is worth. If you are looking for the foundation to start your next thing, you can browse active projects. And if you want the longer thinking on what we are doing and why, the story is here.

The graveyard is not going anywhere. The only question is whether anything inside it gets to live again.